Evaluation of Machine Learning Algorithm for Copy Move Forgery Detection

EasyChair Preprint no. 11523

Abstract

Recently the need for verifying the authenticity of digital images continues to grow, extensive research efforts are dedicated to exploring techniques for detecting image forgeries. Among the prevalent forms of digital tampering, copy-move forgery (CMF) stands out as a widely studied challenge. This manipulation involves duplicating a portion of an image and subsequently pasting it either within the same image or onto a different one. The consequence of such forgery is the obfuscation of the original image content. This study, presents a comprehensive evaluation of four machine learning algorithms, namely k-Nearest Neighbours (kNN), Regression (LR), Naïve Bayes (NB), and Convolutional Neural Network (CNN), for the task of detecting copy-move forgery in images. Our research leverages the CoMoFoD dataset, a widely recognized benchmark for image forensics, to conduct a rigorous assessment of these algorithms.Through analysis, we reveal the strengths and weaknesses of each algorithm in addressing the challenges posed by copy-move forgery detection (CMFD).

Keyphrases: CNN, Copy-move forgery detection, KNN, LR, Machine Learning Algorithms, NB

@Booklet{EasyChair:11523,
  author = {Vijay Bharti Punia and Rohini Goel},
  title = {Evaluation of Machine Learning Algorithm for Copy Move Forgery Detection},
  howpublished = {EasyChair Preprint no. 11523},

  year = {EasyChair, 2023}}